Prediction of Flows With Strong

V. De Henau Curvature and Pressure Gradient

G. D. Raithby Using the k—e Turbulence Model
B. E. Thompson1 The k-e turbulence model of Launder and Spalding has found widespread applica-
tion in the computation offluid flows. Under conditions of strong adverse pressure
gradient or strong streamline curvature, the accuracy of the model is known to
Department of Mechanical Engineering,
University of Waterloo,
decrease. The present paper adds to the available case studies involving these types
Waterloo, Ontario,
offlows by providing predictions for two problems. In the first problem, Simpson's
Canada, N2L 3G1 flow, an adverse pressure gradient induces the flow along a planar surface to
separate. The second problem is unseparated flow over an airfoil. In addition to
predictions using the standard k - t model, results are reported from another k - e
model by Hanjalic and Launder, that has been modified to account for adverse
pressure gradients.

1 Introduction
The critical uncertainty in the numerical simulation of tur-
bulent shear flows is the estimation of the Reynolds stresses.
Models in current use may be as simple as explicit algebraic
relations that connect the stresses to the mean flow variables,
or as complex as a set of differential transport equations with
one equation for each stress. The k — t model of Launder and
Spalding [1] has emerged as a standard because it has been
quite successful for predicting a wide variety of flows, yet is
reasonably simple.
Some authors have, however, reported serious deficiencies
in the A: —e model for particular flows. For flows with strong
streamline curvature, Leschziner and Rodi [2] found that the
measured anisotropy of the normal stresses and the magnitude
of the shear stresses are not well predicted by the k—t model.
For flows in a strong adverse pressure gradient, Rodi and
Scheuerer [3] found that the predicted dissipation is too small,
with the result that turbulent stresses are too large.
Some attempts have been made to modify the k — t model
and improve its performance for such flows. Leschziner and
Rodi [2], and Pourahmadi and Humphrey [4] suggested mak-
ing the C^ "constant" in the model dependent on the flow
curvature. This improved the predictions for jet flows, but
created convergence difficulties [5] for an internal flow. In
their study of flows with strong adverse pressure gradients,
Hanjalic and Launder [6] found that a higher production of
dissipation was required to correctly predict the turbulence
levels for cases where the mean flow is irrotational. This led to
a proposal [6, 3] for increasing the coefficient that multiplies
the production of turbulent kinetic energy by normal stresses
in the e equation.
The present study was undertaken to examine the per-
Current Address: Scientific Research Associates, Glastonbury, Conn.
Contributed by the Fluids Engineering Division for publication in the JOUR-
NAL OF FLUIDS ENGINEERING. Manuscript received by the Fluids Engineering
Division December 12, 1988.
formance of the k—t model, and to evaluate the improvement
of the performance obtained by the Hanjalic and Launder [6]
modification2, for two flows. Both cases were chosen because
of their relevance to airfoil flows. The first case, Simpson's
problem [7, 8, 9, 10], was flow along a wind tunnel wall with
an imposed pressure distribution that approximated the
pressure on an airfoil. The second case was the flow over a
supercritical airfoil measured by Nakayama [11]. Simpson's
flow experiences a strong adverse pressure gradient but no
boundary-induced streamline curvature (internal streamline
curvature does result from flow separation). Nakayama's flow
is subject to both strong adverse pressure gradients and strong
streamline curvature.
To choose an appropriate turbulence model for a computa-
tional fluid dynamics code, the analyst must know the
strengths and weaknesses of all candidate models. This
knowledge is gained through comparisons of predictions with
data for a variety of flows. The aim of this paper is to add two
case studies to the literature that will assist potential users to
understand more completely the capabilities of the standard
k — t model, and the potential benefits of implementing the
Hanjalic and Launder [6] modification.
2 Formulation and Solution Method
2.1 Equations of Motion. For a turbulent, incompressi-
ble, two-dimensional flow, the time-averaged differential
equations of motion, written for a general orthogonal coor-
dinate system, are [12]:
Conservation of mass
{ph2U)+—(PhiV) =0 (1)
h,h' 7• [ • Idx dx-,
The model has been generalized in the present paper to make it coordinate
system independent.

40/Vol. 112, MARCH 1990 Transactions of the ASME

Copyright © 1990 by ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab
 Conservation of m o m e n t u m in the x, direction 1 dU dh.
- ( • (7)
1 T dd dd 1 hi 3x, hxh2 dx2
-~(ph2UU)
— \~-{ph 2UU)+—- + — (phl VU)
dt h,h,
xn2 L3x,
- 3x, dx22
dx J 1 dV U dh-,
- ( • (8)
1 dp i r a , d i h2 dx2 hxh2 dx, '
h, dx. 1 / 1 dV 1 dU V dh2 U 8hi
•+ -
1 dh-, dhx Bh, dh, -/i, dx, h, dx. i h-, dx. h,h, dx2 )
hxh2 •tL"pVV- 3x,
— pUV-
"" 3x 2
0X ~' z 3x
dx.2
•-<722"
" " 3x,
(2)
(9)
Conservation of m o m e n t u m in the x 2 direction 2.2 Turbulence Modelling. T h e standard k — e turbu
lence model calculates the turbulent viscosity fi, as [1]
3 a
1(PV)+TJT2 dx (ph2UV) + dX;
-{phxVV)\
/V
pk2
(10)
1 dP 1 r 3 3
where e is the rate of dissipation a n d CM is taken as a constant.
h, dx-, /),/!, Ldx, 3x, The transport equations for the turbulence kinetic energy a n d
the dissipation rate in general orthogonal coordinates can be
1 r 3/i, 3/i, 3/i 2 3/i,
— \pUU-—-—pUV—^- + o2l —a, (3) written as [13]:
h l"2 9x, dx. dx. 3x, Transport of turbulence kinetic energy k
h L
In the equations above, U and K a r e t h e mean velocity in the 3 1
X, and x 2 directions, respectively, p is the density a n d h{ a n d — (ph2Uk)+ — (phlVk)
-dt5 7 ( 0 * ) + Thx7hr2
dx, dx 2
h2 are the metrics. T h e local effective mean pressure P is
defined as
hxh2 L3x, V hx V ok ) dXi )
P*+—pk P=
3 9k
where P* is the mean pressure relative to the hydrostatic (11)
i 2 \ h 2 \ ak ) u*2
pressure and k is the turbulence kinetic energy. T h e turbulent
Transport of rate of dissipation e
viscosity approach relates the stresses a,y t o the strain rates etj
through the relations 3
, s l
( p / l 2 t / e ) + - - ( ? / * , Ke)
dt hxh2 Lax,
on=2(n + nt)en (4)
1 8e
o22 = 2{n + H,)e22 (5)
h,h. Ldx, \ /I, V ffe / 3x, /

on = 2(ji + n,)en (6)

where /x, is the turbulent viscosity. For incompressible flows, + Pt-Ce2peT (12)
dx, V h-, \ a, / dx, /

Nomenclature

c airfoil chord length y+ = nondimensional distance normal

^-ji> el' constants for turbulence model from a wall
fluid strain d = boundary layer thickness
wall turbulence model constant, 5* = displacement thickness
equation (21) e = rate of dissipation of turbulence
Ai. A2 = metrics in x , a n d x 2 directions, energy
respectively K, K, = wall turbulence model constants
/i s , h„ = metrics in s a n d n directions, = laminar and turbulent viscosities
/*. A1/
respectively ", = turbulent kinematic viscosity
H = height of calculation domain = density
P
k = turbulence kinetic energy a = fluid stress
L = length of calculation domain a = turbulent Prandtl number for k
k> °e
P* = pressure and e, respectively
P = P* + 2Apk T = inverse time scale = e/k
Pk,Pe = production of turbulence energy = wall shear stress
T,v
and of dissipation, respectively
Re x = Reynolds number based on Subscripts
characteristic length x 1,2 = denoting coordinate directions x,
s,_n = streamline coordinate directions and x 2 , respectively
, vv, uv = turbulent Reynolds stresses s, n = denoting coordinate directions s
U,V = mean velocity components in the and «, respectively
x, a n d x 2 directions, respectively t = turbulent
u„ u„ = mean velocity components in the s 00 = referring to freestream condition
and n directions, respectively
ur = shear velocity Acronyms
u+ = velocity non-dimensionalized by SKE = standard k — e turbulence model
shear velocity MKE = modified A:—g turbulence model
generalized coordinates directions with C e 3 = 5.50

Journal of Fluids Engineering MARCH 1990, Vol. 1 1 2 / 4 1

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab

 The empirical constants in the standard k — e model are given
the following values [1]: C „ = 0 . 0 9 , ak=l.O, ffE = 1.3, $=constant
Ce[ = 1.45 and Ct2 = 1.92. r is an inverse time scale defined as

The production of turbulence kinetic energy Pk is expressed

in terms of Reynolds stresses and mean flow strain rates as
Pk= -puv(2en)-puuen -pvve22 (13)
and the "production" of dissipation Pt is approximated by
--C,,Pt (14)
To permit a different sensitivity of P £ to the normal stresses,
Hanjalic and Launder [6] substituted equation (13) into equa-
tion (14), eliminated e22 in favor of en using mass conserva- X
tion, and then changed the coefficient multiplying the normal Fig. 1 Streamline coordinate system
stresses from Cel to Ce3 to obtain the following equation.
Pe=r[- CeXpuv{2en)- C^p(uu-vv)eu\ (15)
To augment the production of dissipation by normal strains, UT = C]/'kl/2
Cci is positive and greater than C e l . A value of Cf3 of 4.44 was For the viscous and buffer layers (y+ < 30), Rannie [15] sug-
recommended by Hanjalic and Launder but Rodi and gested that
Scheuerer [3] used 2.5 for high adverse pressure gradient 1
flows. C/+ -tanh(K,/ + ) (20)
Hanjalic and Launder [6] use the following expression for
(uu — vv), where K, =0.0688 and U+ = U/UT. In the fully turbulent layer
(uu-~vv)= 0.33k (16) region ( y + > 3 0 ) , the standard logarithm velocity profile is
adopted [1], i.e.,
The value of 0.33 is in agreement with experimental results for
boundary layer flows in adverse pressure gradients [14]. Equa- < J + = — ln(£> + ) (21)
tion (16) is different from the expression K

1 dU dh. where K and E are the log-law constants equal to 0.41 and 9.0
(uu — vv) = — • (for a smooth wall), respectively [1].
P h{ dxx h1h2 dx2
Expressions for the shear stress at the wall are obtained
which follows from the turbulent viscosity definition of from equations (20) and (21), assuming that for the region
stresses. For a fully developed flow between parallel plates, very near the wall the pressure gradient has no influence on the
this equation yields (uu — vv) =0, which is contrary to flow [16] and the flow is locally one-dimensional. The expres-
measurements. Therefore, the present study has followed sions are
Hanjalic and Launder by using equation (16). for^<_vL
To apply this model to a flow not aligned with the xt direc-
:\MklP/2K
tion (assumed in equation (16)), the local s — n coordinate (22)
system in Fig. 1 is introduced in a manner similar to that pro- tanh(K,.Vp)
posed by Leschziner and Rodi [2]. The expression for Pk in for yp~>y\ +
this coordinate system is
pUpC^k1?
Pk = 4n,(esn)2-0.33pkess (17) (23)
where ess and es„ are the normal and shear strains respectively. The subscript P refers to the node in the flow that lies adjacent
These are defined in Appendix A. to the wall while/1 + is the value of y+ at which equations (22)
Equation (17) is used to calculate the production of tur- and (23) give the same value of r„.
bulence kinetic energy in the ^-equation (equation (11)) as well
as in the e-equation (equation (12)) where P e is expressed as Wall Boundary Condition for k and e. At the wall,
dk/dy = 0. In addition, assuming that the dominant term in the
Pc =r[CelPk - ( C d - C€l)0.33PkeJ (18) near wall production is the shear stress, the following can be
The derivation of equations (17) and (18) is given in Appendix derived from equations (20) and (21)
A. for^+<^2 +
2.3 Boundary Conditions. This section deals with the n P2U\ tanh2(Kl>< + )
(24)
application of the boundary conditions for flow along a tur- \x. cosb2(Kiy + )
bulent wall. fory+>y2*
Wall Boundary Conditions on Velocity. The velocity nor- l
Ut 1
mal to a solid wall is set to zero. The boundary condition for (25)
Ky +
the tangential velocity is specified by a shear stress based on a /*
wall function approach described below. y2 + is the value of y+ at which equations (24) and (25) give
In the derivation of the wall function, y+ is taken as the the same value of Pk. The production of k for a given node
non-dimensional distance normal to the wall that lies immediately adjacent to the wall is obtained by in-
pyUT tegrating equations (22) and (23) over the y+ region occupied
(19) by the node.
r Based on the assumption that the flow is in local
If there is an equilibrium layer near the surface, in which the equilibrium in the near wall region (pe = Pk), the variation of e
shear stress is nearly constant and the production of k is in with distance normal to the wall is simply expressed as
balance with its dissipation, then f o r 5 < j + <yl +

42/Vol. 112, MARCH 1990 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab

 Air Properties
0 = 1.164 Kg/m'
I = 7.495 m (1= 1.824xI0-'(V.«/m
H = 0.355 m
Separation Point
C
Fig. 3 Streamlines for turbulent separating boundary layer flow
T^,,l'l
" .1' ' " " ' . ' (

111 lll'.cl'.l.lljL' (
I J

Fig. 2 Wind tunnel geometry and fluid properties for the turbulent
separating boundary layer flow problem

Table 1 Experimental parameters of the mean flow

development
x 6 8*
(m/s) (cm) (cm)
(m)
2.778 17.038 6.066 1.601
2.858 16.599 6.760 —
2.934 16.185 7.219 2.317
3.123 15.228 8.828 —
3.273 14.661 11.877 5.245
3.448 14.158 14.818 7.889
3.604 13.868 17.191 10.191
3.896 13.594 25.568 15.888
3.996 13.517 27.861 Fig. 4 Bottom wall shear stress distribution for turbulent separating
— boundary layer flow (60 x 35 grid).
Experimental Simpson: o; SKE: ; MKE:

pU\ tanh 2 (K,^ + )

e =
p. cosh2 (K .J* + )
iory+ > y2+
P&r 1 tions reported in this paper avoided this problem [18, 19].
In the viscous sublayer (y+ <5), measurements indicate that
the dissipation rate is constant [1]. The level of e for that
region is therefore given by the level of e at y+ = 5, that is
fory+ < 5
2 max
PU\ tanh (5K,)
H cosh2(5K,)
Equations (26) to (28) replace the transport equation for the
dissipation rate (equation (12)) at the nodes that lie immediate-
ly adjacent to the wall.
It is noted here that the treatment of the wall boundary con-
ditions for the region y+ <30 is different from the standard
wall function method [1]. When applied to the fully developed
turbulent flow between parallel plates of Laufer [17], the stan-
dard wall model predicts a decrease in k near the wall which is
more rapid than the one observed experimentally. The present
wall model yields the correct trend for k near the wall and, for
this reason, is adopted in this paper.
Summary: From known U and k fields (obtained from a
previous iteration or from an initial guess), the values of y£
are evaluated. Based on these values, the wall shear stress, as
well as the production of k and the dissipation level at the
wall, are calculated and used as boundary values in the ap-
propriate equations.
2.4 Numerical Model and Solution Procedure. Informa-
tion on the performance of the turbulence model presented in
the previous section will be obtained by comparing the solu-
tion of the equations to measurements. For this information'
to be reliable, it is necessary that the numerical errors in the
solution do not contaminate the comparison.
The discretization method used in the present study was ex-
tensively tested [18, 19] for a variety of external flows for
which benchmark solutions were available. Furthermore, grid
independence studies were conducted on the solutions to the
two flows described in this paper. Results of these studies are
discussed in the followng section.
It may be of interest here to alert analysts to the serious
(26)
pressure errors that occur when upwind differences are used.
These, of course, contaminate the velocity solution for exter-
nal flows [20, 19]. The discretization used to obtain the solu-
(271
Solutions to the non-linear equation set were obtained from
successive iterations on a linear form of the equations. Itera-
tions were performed until the following criterion was met
tj,n+l _tj>" i
T rL"+l
< 0.00001
T
mi n
(28) where $ represents k, e and the pressure P and n is the itera-
tion number. This criterion is close to the round-off limit of a
VAX 780 computer.
3 Numerical Results
3.1 Turbulent Separating Boundary Layer Flow
Description of the Problem. Simpson et al. [7-10]
measured the two-dimensional flow in a wind tunnel with the
cross-sectional shape shown in Fig. 2. The goal was to impress
a pressure on the lower planar surface similar to that en-
countered on an airfoil. To avoid separation on the upper sur-
face of the wind tunnel, mass was withdrawn through slots at
locations A, B, and C. Separation was observed on the lower
surface at about 40 percent of the distance to C from location
B (between 3.10m to 3.45m from the inlet). The area of the
tunnel was reduced past C to close off this recirculation
bubble.
The measurements presented by Simpson et al. were nor-
malized with respect to a velocity scale U„ (x) and length
scales of 5 (x) and 8* (x). The values of these quantities are re-
quired at the x-locations at which comparisons between the
measurements and the predictions are made. The relevant
quantities are reproduced in Table 1.
Implementation of the Numerical Model. The numerical
model for the problem was run using the kind of orthogonal
grid shown in Fig. 2. The results presented here were obtained
with a grid of density (60x35).
In accordance with the experiment, the inlet velocity was
prescribed as uniform (U= 15.057 m/s) in the x direction and
zero (V= 0) in thej> direction, k was specified to correspond to
a uniform isotropic turbulence intensity of 0.1 percent.

Journal of Fluids Engineering MARCH 1990, Vol. 112/43

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab

 1.2 n 1 1 1—
x = 2.934m
1.0

0.6

0.4

0.2

0.0 <»*
0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 0.0 0.2 0.4 0.6 0.8 1.0

Fig. 5 Comparison of experimental and numerical mean velocity pro-

files (60 x 35 grid).
Experimental Simpson: o ; SKE: ; NIKE:

X = 3.996TI

0.00 0.010.00 0.010.00 0.010.00 0.01 0.02 0.03

Fig. 6 Comparison of experimental and numerical turbulence kinetic

energy profiles (60 x 35 grid).
Experimental Simpson: o ; SKE: ; MKE:

0.0
0.000 0.006 0.000 0.006 0.000 0.006 0.000 0.006

Fig. 7 Comparison of experimental and numerical turbulent kinematic

viscosity profiles (60 x 35 grid).
Experimental Simpson: o; SKE: ; MKE:

Following the recommendation of Cebeci et al. [21], an inlet
dissipation rate of e = k3/2/(0.005H) was prescribed, were His
the wind tunnel height at the inlet. The wind tunnel walls were
taken as hydraulically smooth. Separation on the upper sur-
face was avoided in the computer model by withdrawing mass
through the wall between B and C at a uniform rate which
matched the measured total withdrawal. This produced
velocities normal to the wall of about 0.7 percent of the inlet
velocity.
Numerical solutions were obtained with the standard k — e
model (SKE), and with the modified model (MKE), in which
Ce3 = 5.50.
Presentation and Discussion of the Results. The SKE
model failed to predict separation on the lower surface. The
MKE model predicted separation at *~3.4m, which lies
within the range of observed separation. To determine the sen-
sitivity of the gross features of the flow to Ce3, solutions were
obtained for several different values of Cei. Flow separation
was found for Ce3>2.50. As Ce3 was increased above this

44/Vol. 112, MARCH 1990 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab

 1 I

Experimental x = 3.896m: o ir
60x35 grid : 1
1.0
60x21 grid : IP
95x21 grid : I

0.8 / -
0 J
O y//
TS 0.6 /v -
° /y
0.4
0
/? _
o fl
0.2 - 0 / - 0.0 4.0 8.0 0.0 4.0 B.O 0.0 4.0 8.0 0.0 4.0 8.0
0o / (BU/t/j;) x 1000
/
%JL\ I i i i i Fig. 10(a) uu Profiles for supercritical airfoil at 4 deg incidence.
0.0 1 1

Experimental Nakayama: o;SKE: ;MKE:

- 0 . 2 0 . 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 for x/C< 1.000, y' = y; for X / C > 1 . 0 0 0 , y' = y-yc
6
U/U„ Re = 1.2x10
Fig. 8 Comparison of velocity profiles obtained with MKE for various
grid sizes at a station located in the recirculating flow region.

C y

Fig. 9 Super critical airfoil profile and coordinates definition

value, the separation bubble increased. For C t3 >5.50, con- 0.0 4.0 8.0 0.0 4.0 B.O 0.0 4.0 8.0 0.0 4.0 8.0
verged solutions could not be obtained. The streamline for the (un/ul) x 1000

MKE model (Ce3 =5.50) are shown in Fig. 3.
Comparisons of the wall stresses are shown in Fig. 4. The
SKE model degrades in accuracy in the adverse pressure gra-
dient zone as the separation point is approached and fails to
change sign (i.e., no separation). The MKE model departs
from the SKE predictions in the adverse pressure gradient
region and agrees much better with the measurements near
separation. The accurate prediction of the location of the
separation point already discussed, is reflected in this plot by a
change in the sign of TW at the correct ^-location.
Profiles of the U-velocity comparison is shown in Fig. 5 at
four stations, two of which are downstream of separation.
The overall level of agreement is much better with the MKE
model. The height of the backflow (£/<0) region within the
separation bubble, and the velocity defect within the boundary
layer (y/5<l), are significantly underpredicted by both
models.
The turbulent kinetic energy and turbulent viscosities are
compared in Figs. 6 and 7, respectively. There is some im-
provement in the k profiles (Fig. 6) predicted by the MKE
model, compared to SKE, only in the separated flow region,
but there is generally poor agreement with the measurements
in this region: the predicted level of k is too small, and the
predicted region of high k is confined too close to the wall.
The turbulent viscosity (Fig. 7) is greatly overpredicted by the
SKE model upstream of the separation (this prevents the flow
from separating). It is surprising that the vt profiles predicted
downstream of separation are almost identical for the SKE
and MKE models, even though flow separation has occurred
with the MKE model.
Influence of Grid Refinement of the Predictions. Calcula-
tions of the Simpson flow were performed on various grid
sizes in order to investigate the influence of grid refinement on
the accuracy of the predictions. Figure 8 illustrates the velocity
profiles obtained with MKE at a station located in the recir-
culating flow region for (60 x 35), (60 x 21) and (95 x 21) grids.
Although a small effect of grid size on the predictions can be
Fig. 10(b) mi Profiles for supercritical airfoil at A deg incidence.
Experimental Nakayama: o;SKE: ; MKE:
for x/C<1.000, y* = y; for x/C>1.000, y* = y - y c
Re = 1.2x10 6
observed, it is not responsible for the important discrepancies
between the predictions and the experimental data. Similar
conclusions can be drawn for the predictions of k and v,.
Summary. The MKE model provides much better agree-
ment with the data than the SKE model for Simpson's
problem. There are, however, very significant discrepancies
between results from this model and the data, particularly in
the separated flow region.
3.2 Turbulent Flow Past a Supercritical Airfoil With 4°
Angle of Attack at Rec = 1.2 X106
Description of the Problem. Nakayama [11] has provided
measurements of flow over the supercritical airfoil shown in
Fig. 9, with an angle of attack a = 4°. Measurements were ob-
tained in a wind tunnel with walls located at 0.79C above and
below the airfoil. The inlet velocity was 30 m/s, so that the
Reynolds number based on the chord length C was 1.2x 106,
and the turbulence intensity was 0.02 percent. The coordinate
system in which the data were reported is also shown in Fig. 9.
Implementation of the Numerical Model. A 95 x 40 or-
thogonal grid extended 10 chord lengths upstream and
downstream of the airfoil. The upper and lower boundaries
were placed at the location of the wind tunnel wall, and the
grid conformed to the airfoil surface. Of the 95 control
volumes in the x, direction, 33 were clustered on both the top
and the bottom surfaces of the airfoil. Such a grid was found
to provide adequate resolution.
The inlet velocity and kinetic energy (isotropic turbulence
assumed) were matched with the experiment. The inlet dissipa-
tion was selected to make the turbulent viscosity 3 to 4 orders
of magnitude smaller than its maximum value near the airfoil.
Zero normal velocity and slip conditions were applied on the

Journal of Fluids Engineering MARCH 1990, Vol. 112/45

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab

 0.06
_1 ! ! ! ! ! ! , , , T_
c
| x / C = 0.893 W = 0.970 a \ * / c = 1.000 „ r
/ C = 1-010
0.04

0.02

O
«-- o.oo

-0.02

-0.04

-0.06
0.0 4.0 S.O 0.0 4.0 8.0 0.0 4.0 8.0 0.0 4.0 8.0 -6.0 -3.0 0.0 3.0 0.0 3.0 0.0 3.0 0.0 3.0 6.0
(vB/U*) x 1000 W £ ) x 1000

Fig. 11(a) vv Profiles for supercritical airfoil at 4 deg incidence. Fig. 12(a) uv Profiles for supercritical airfoil at 4 deg incidence.
Experimental Nakayama: o ; SKE: ; MKE: Experimental Nakayama: o ; SKE: ; MKE:
for x/C < 1.000, y* = y; for x/C>1.000, y* = y-yc for x/C< 1.000, y* = y; for x/C> 1.000, y* = y-yc
Re = 1.2x10 6 Re = 1.2x10 6

x/C =1.020 i / C = 1.050 z / C = 1.200 i/C=1.400

0.04

0.02

J i 0.00 iio.oo

-0.02

-0.04

-0.06 -6.0 -3.0 0.0 3.0 6.0-3.0 0.0 3.0 0.0 3.0 0.0 3.0
0.0 4.0 8.0 0.0 4.0 B.O 0.0 4.0 8.0 0.0 4.0 8.0
(mi/t/^) x 1000
(CO/D^) x 1000
Fig. 11(b) vv Profiles for supercritical airfoil at 4 deg incidence. Fig. 12(b) uv Profiles for supercritical airfoil at 4 deg incidence.
Experimental Nakayama: o ; SKE: ; MKE: Experimental Nakayama: o;SKE: ;MKE:
forx/C<1.000, y* = y; for x/C>1.000, y* = y-yc forx/C<1.000, y* = y; for x/C>1.000, y' = y-yc
6
Re = 1.2x10 6 Re = 1.2x10

wind tunnel walls. With this boundary condition, the flow far
away from the airfoil is nearly inviscid, but the disturbance of
the pressure around the airfoil by the wall is accounted for in
the model. It would have been too expensive, computational-
ly, to resolve the viscous boundary layer growth along the tun-
nel wall.
Comparisons of turbulent stress measurements and predic-
tions are presented along j'-profiles at several x/C stations be-
tween 0.893 and 1.40. x/C< 1.0 denotes stations on the airfoil
and, at such stations, the ^-coordinate lies normal to the sur-
face of the airfoil (see Fig. 9).
Presentation and Discussion of the Results. Figures 10(a)
and 10(b) compare the measured and predicted uu stress. For
stations on the airfoil, the SKE model provides more accurate
predictions, both above (y/C>0) and below the airfoil. Just
beyond the trailing edge (x/C= 1.02), both models give similar
predictions, but further into the wake, the MKE stresses decay
too quickly. Neither model however provides good agreement.
For example, neither model predicts the double peak of uu in
the wake, that probably results from the advection of uu by
the streams from above and below__the airfoil.
The levels of the normal stress vv are in better agreement
with the measurements (Figs. 11(a), \\(b)), but the double
peak peak that exist at the trailing edge again disappears
within 1 percent of the chord length. Clearly, the eddy
viscosity approach depends too strongly on local effects, with
too little memory of upstream conditions. For this vv stress
component, the MKE model predicts somewhat better levels
than the SKE model.
Of particular interest are the uv stresses shown in Figs.
12(a) and 12(b). The MKE model is significantly bet-
ter than the SKE model on the airfoil (Fig. 12(a)), while the
reverse is true in the wake (Fig. 12(b)). The magnitude, and
the decay rate behind the airfoil, of the shear stresses is better
predicted than the normal stresses.
Summary. For Nakayama's problem, in which both
strong streamline curvature and adverse pressure gradients oc-
cur, some stresses are better predicted by the MKE model and
some by the SKE model. The MKE model tends to be better
adjacent to the airfoil and the SKE is better in the wake. While
the general stress levels are reasonably well predicted, the pro-
files are incorrect in the wake.
4 Conclusions
The performance of the standard k-e model (SKE) as well
as that of a modified k — e model (MKE), is evaluated for two
case studies that involve strong adverse pressure gradients
and, for one of the cases, strong boundary induced streamline
curvature. The purpose of the modification in the MKE model
is to sensitize the dissipation equation to the normal stresses.
For the first problem, involving a sufficiently strong adverse
pressure gradient to induce separation, but with no boundary
induced curvature, the modified (MKE) model induced
separation while the standard k — e (SKE) model did not. The
failure to predict such an important flow feature represents a
major deficiency of the standard k—e model.
The second problem, of flow over a supercritical airfoil at 4
deg angle of attack, has a strong boundary curvature and in-
duces strong adverse pressure gradients on the suction surface.
In this case, the stresses predicted by the MKE model were
slightly more accurate for the flow near the surface, but the
SKE model provided more accurate stresses in the wake. The
level and wake decay rate of the stresses were roughly correct

46/Vol. 112, MARCH 1990 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab

 for this problem. The profile shapes, especially in the wake, as the mean velocity along a streamline and U„, the velocity
were poorly predicted. normal to a streamline. Since, by definition, no mass flows
The results presented in this paper underline the poor per- across a streamline, Us and U„ are given by
formance of the SKE model in predicting flows with adverse
pressure gradient and strong streamline curvature. As ex-
pected, the MKE model offers an improvement over the SKE
model for the case of the strong adverse pressure gradient flow L7„ = 0
but performs poorly for the curved flow. This is not surprising Therefore, the continuity equation for a stream tube is
considering that the MKE model has not been enhanced to
treat such flows. It was also found that the predicted turbulent 1 d
-(phnUs)=0 (29)
stresses do not have sufficient memory of upstream condi- hsh„ ds
tions. This underlines the fundamental inadequacy of the tur- where hs and h„ are the metrics in the ^ and n directions,
bulent viscosity formulation. respectively.
The strains are expressed as
References
1 Launder, B. E., and Spalding, D. B., "The Numerical Computation of
(30)
hs ds
Turbulent Flows," Comp. Methods Appl. Mech. Eng., Vol. 3, 1974, pp.
269-289.
2 Leschziner, M. A., and Rodi, W., "Calculation of Annular and Twin
U, dh„
*~nn (31)
Parallel Jets Using Various Discretization Schemes and Turbulence-Model hnhs ds
Variations," ASME JOURNAL OF FLUIDS ENGINEERING, Vol. 103, 1981, pp.
352-360.
3 Rodi, W., and Scheuerer, G., "Scrutinizing the k — e Turbulence Model
Under Adverse Pressure Gradient Conditions," ASME JOURNAL OF FLUIDS
ENGINEERING, Vol. 108, 1986, pp. 174-179.
4 Pourahmadi, F., and Humphrey, J. A. C , "Prediction of Curved Chan-
1
- ! - [ • - h„
dU,
dn
U< dh< 1
hsh„ dn "~2~ vtim™
By expanding equation (29), it can be easily shown that
nel Flow with an Extended k — e Model of Turbulence," AIAA Journal, Vol. 21,
1983, pp. 1365-1373. ess=-e,m (33)
5 Hackman, L. P., " A Numerical Study of the Turbulent Flow Over a
Backward Facing Step Using a Two Equation Turbulence Model," Ph.D.
The Reynolds stresses are related to the mean flow gradients
thesis, University of Waterloo, Waterloo, Ontario, Canada, 1982. using the Boussinesq's approximation and are given by
6 Hanjalic, K., and Launder, B. E., "Sensitizing the Dissipation Equation
to Irrotational Strains," ASME JOURNAL OF FLUIDS ENOINEERING, Vol. 102, - pusus = 2ixtess——pk (34)
1980, pp. 34-40.
7 Simpson, R. L., Chew, Y.-T., and Shivaprasad, B. G., "The Structure of
a Separating Turbulent Boundary Layer. Part 1. Mean Flow and Reynolds
Stresses," J. Fluid Mech., Vol. 113, 1981, pp. 23-51. -pu„u„=2iilem——pk (35)
8 Simpson, R. L., Chew, Y.-T., and Shivaprasad, B. G., "The Structure of
a Separating Turbulent Boundary Layer. Part 2. Higher-Order Turbulence
Results," J. Fluid Mech., Vol. 113, 1981, pp. 53-73. - pusu„ = 2jx,es„ (36)
9 Shiloh, K., Shivaprasad, B. G., and Simpson, R. L., "The Structure of a
Separating Turbulent Boundary Layer. Part 3. Transverse Velocity
Measurements," J. Fluid Mech., Vol. 113, 1981, pp. 75-90.
10 Simpson, R. L., Shivaprasad, B. G., and Chew, Y.-T., "The Structure of A.l Derivation of Pk
a Separating Turbulent Boundary Layer. Part 4. Effects of Periodic Free- The expression of Pk is s — n coordinates is given by
Stream Unsteadiness," J. Fluid Mech., Vol. 127, 1983, pp. 219-261.
p
11 Nakayama, A., "Characteristics of the Flow Around Conventional and k = -pusu„ (2es„)-pu„u„e„„ ~pususess
Super Critical Airfoils," / . Fluid Mech., Vol. 160, 1985, pp. 155-179.
12 Raithby, G. D., Galpin, P. F., and Van Doormal, J. P., "Prediction of From equation (33), Pk can be written as
Heat and Fluid Flow in Complex Geometries Using General Orthogonal Coor- p
dinates," Numer. Heat Transfer, Vol. 9, 1986, pp. 125-142. k = -f>usu„ (2esn) -p(usus-u„un)ess (37)
13 Hackman, L. P., Raithby, G. D., and Strong, A. B., "Numerical Predic- Following the proposal of Hanjalic and Launder [6],
tion of Flows Over Backward-Facing Steps," Int. J. Num. Meth. Fluids, Vol. 4,
1984, pp. 711-724.
(usus — u„un) is related to the turbulence kinetic energy k by
14 Simpson, R. L., "Characteristics of a Separating Incompressible Tur- Pk = -PUsun (2es„)-0.33pkess
bulent Boundary Layer," AGARD CP 168, paper 14, 1975.
15 Rannie, W. D., "Heat Transfer in Turbulent Shear Flow," J. Aeronaut. or, by substituting the equation for the Reynolds shear stress
Sci., Vol. 23, 1956, pp. 485-489. (equation (36)), we obtain
16 Lewis, J. P., and Pletcher, R. H., " A Boundary-Layer Computational
Model for Predicting the Flow and Heat Transfer in Sudden Expansions," Pk = 4n,(esn)2-0.33pkess (38)
Technical Report No. HTL-26, CFD-14, ISU-ERI-Ames-87018, Engrg.
Research Institute, Iowa State University, Ames, 1986.
17 Laufer, J., "Some Recent Measurements in a Two Dimensional Turbulent A.2 Derivation of Pt
Channel Flow," Journal of Aeronautical Sciences, 1950, pp. 277-287.
18 De Henau, V., " A Numerical Study of the Turbulent Flow Around Air- In the standard k — e model, the generation of dissipation is
foils," M.A.Sc. thesis, University of Waterloo, Waterloo, Ontario, Canada, given by
1987.
19 De Henau, V., Raithby, G. D., and Thompson, B. E., "A Total Pressure P*=CtlPkT
Correction for Upstream Weighted Schemes," Int. J. Num. Meth. Fluids, Vol.
9, 1989, pp. 855-864. By decomposing Pk into its components parts (equation (38)),
20 Castro, I. P., Cliffe, K. A., and Norgett, M. J., "Numerical Predictions Hanjalic and Launder [6] defined Pc as
of the Laminar Flow Over a Normal Flat Plate," Int. J. Num. Meth. Fluids,
Vol. 2, 1982, pp. 61-88. P e =4C e l TMe s „) 2 -0.33C t3 T P /ce JS (39)
21 Cebeci, T., Khalil, E. E., and Whitelaw, J. H., "Calculation of Separated
Boundary-Layer Flows," AIAA J., Vol. 17, No. 12, 1979, pp. 1291-1292.
Equation (39) can be rewritten as
Pt=4CtlTn,(esn)2-0.33CtlTPkess
A P P E N D I X A
-0.33Ce3Tpkess + 0.33CelTpkess
A Derivation of Pk and Pt in the Streamline Coordinate
System or, more conveniently
Figure 1 illustrates the s — n coordinate system. Us is defined Pt = r[CclPk - (C£3 - C el )0.33pfc?J (40)

Journal of Fluids Engineering MARCH 1990, Vol. 112/47

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/journals/jfega4/27047/ on 03/23/2017 Terms of Use: http://www.asme.org/ab